Word 表格汇总到 Excel

lijt16

Word 表格汇总到Excel
Word 表格汇总到Excel
Word 表格汇总到Excel
具体使用看Word 表格汇总 Excel成品。

别再发看不懂了，看不懂的麻烦Ctrl+W关闭此贴，
别浪费你我的时间，别发无意义的评论，谢谢。


昨天发了Word 表格汇总 Excel，看评论大家还是需要源码，这样可以根据自己需求进行修改，所以还是上传上来吧，不过本人也是业余的，很多地方有改进的地方。

本来我是用的tkinter，没有图形界面的，想根据自己需要修改的只需要下这个就行了：

[Python] 纯文本查看 复制代码

?

001 002 003 004 005 006 007 008 009 010 011 012 013 014 015 016 017 018 019 020 021 022 023 024 025 026 027 028 029 030 031 032 033 034 035 036 037 038 039 040 041 042 043 044 045 046 047 048 049 050 051 052 053 054 055 056 057 058 059 060 061 062 063 064 065 066 067 068 069 070 071 072 073 074 075 076 077 078 079 080 081 082 083 084 085 086 087 088 089 090 091 092 093 094 095 096 097 098 099 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133

# -*- coding: UTF-8 -*- import tkinter as tk import win32com.client as wc import os, re, glob, time, shutil from tkinter import filedialog from bs4 import BeautifulSoup from pydocx import PyDocX   class Doc2Xlsx(object):         """docstring for Doc2Xlsx"""         def __init__(self):                 super(Doc2Xlsx, self).__init__()           # doc2docx         def doc2docx(self, path, file):                 word = wc.Dispatch("Word.Application")                 doc = word.Documents.Open(path + "/" + file)                 doc.SaveAs(path + "/%sx"%file, 12)  #12代表转换后为docx文件                 doc.Close                 word.Quit()   #结束后台进程             # Del doc         def delDoc(self, path, file):                 time_start=time.time()                 while True:                         time_end=time.time()                         wmi = wc.GetObject("winmgmts:")                         proCodeCov = wmi.ExecQuery('select * from Win32_Process where Name="WINWORD.EXE"')                         if len(proCodeCov) <= 0:                                 os.remove(path + "/" + file)                                 return                         time.sleep(3)                         if time_end - time_start > 20:                                 print('删除 【%s】 文档超时'%file)                                 return             # docx path         def getPath(self):                 root = tk.Tk()                 root.withdraw()                 path = filedialog.askdirectory(title = "选择Word登记表所在目录")                 docs = [i for i in os.listdir(path) if i.endswith(".doc")]                 if docs:                         print("==================Doc2Docx==================")                         for doc in docs:                                 self.doc2docx(path, doc)                                 self.delDoc(path, doc)                                 print("%s 已转换为: %s"%(doc, doc+"x"))                         print("==================Doc2Docx==================\n")                 docxs = [i for i in os.listdir(path) if i.endswith(".docx")]                 return path, docxs             # 根据html tr&td 标签分类         def docxCategory(self, path, docxs):                 docx_cate = {}                   for docx in docxs:                         html = PyDocX.to_html(path + "/%s"%docx)                         soup = BeautifulSoup(html, 'html.parser')                         elem = soup.find_all("tr")                              #每个tr标签对应表格的一行                                              elem1 = re.sub(r"(?<=<td).+?(?=</td>)", "", str(elem))  #去除单元格标签td之间的内容                           if elem1 in docx_cate.keys():                                 e = docx_cate[elem1]                                 e.append(docx)                                 docx_cate[elem1] = e                         else:                                 docx_cate[elem1] = [docx]                   print("==================docxCategory==================")                 for c, docx in enumerate(docx_cate.values()):                         path_cate = path + "/_cate/类别%d"%(c+1)                           if not os.path.exists(path_cate):                                 os.makedirs(path_cate)                           print("类别%d: "%(c+1))                         for i in docx:                                 print(i, end=" ")                                 # shutil.move("%s/%s"%(path, i), "%s/%s"%(path_cate, i))        #移动                                 shutil.copyfile("%s/%s"%(path, i), "%s/%s"%(path_cate, i))    #复制                         print()                 print("==================docxCategory==================")             # docx2xlsx. 模板命名model.docx,与word放置相同文件夹         def docx2xlsx(self, path, docxs):                 docxs.remove("model.docx")                   # 获取模板文件所需数据的位置                 html = PyDocX.to_html(path + "/model.docx")                 soup = BeautifulSoup(html, 'html.parser')                 elem = soup.find_all("tr")                 data = {}                 for row, tr in enumerate(elem):                         if tr.text:                           #该行有需要的数据                                 val = {}                                 units = tr.find_all("td")         #获取需要的单元格序号                                 for unit, td in enumerate(units):                                         if td.text:                                                 val[unit] = td.text                                 data[row] = val                 # 创建xlsx                 xl = wc.Dispatch('Excel.Application')                 xl.Visible = True                 xl.Workbooks.Add()                 xlBook = xl.Workbooks(1)                 xlSheet = xl.Sheets(1)                 # 数据汇总                 try:                         for c, docx in enumerate(docxs):                                 html = PyDocX.to_html(path + "/%s"%docx)                                 soup = BeautifulSoup(html, 'html.parser')                                 elem = soup.find_all("tr")                                 for r in data.keys():                                         for u in data[r].keys():                                                 xlSheet.Cells(c+1, int(data[r][u])).Value = elem[r].find_all("td")[int(u)].text.strip()                 except:                         print(docx,"出错!")                 xlBook.SaveAs(path.replace(r"/", "\\") + "\\result.xlsx")  #实际单\才能正常保存                 # xl.Quit() #关闭     if __name__ == '__main__':         d2x = Doc2Xlsx()         path, docxs = d2x.getPath()         d2x.docx2xlsx(path, docxs)         classNeeded = 0         if classNeeded:                 d2x.docxCategory(path, docxs)

后来因为自己之前就一直想学一下qt模块，看了下pyqt5和pyside2，还是用了后者。
提一下这两者都可以用自带的qtdesigner设计个大概的样式再进行修改的，不过因为第一次用，很多地方不知道怎么用。
首先是转化生成并修改的引用文件，可以直接引用.UI文件，不过因为要打包就转化成py文件了：

[Python] 纯文本查看 复制代码

?

001 002 003 004 005 006 007 008 009 010 011 012 013 014 015 016 017 018 019 020 021 022 023 024 025 026 027 028 029 030 031 032 033 034 035 036 037 038 039 040 041 042 043 044 045 046 047 048 049 050 051 052 053 054 055 056 057 058 059 060 061 062 063 064 065 066 067 068 069 070 071 072 073 074 075 076 077 078 079 080 081 082 083 084 085 086 087 088 089 090 091 092 093 094 095 096 097 098 099 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159

# -*- coding: utf-8 -*- from PySide2.QtCore import * from PySide2.QtGui import * from PySide2.QtWidgets import * import qtawesome   class d2xUI(object):       def setupUi(self, MainWindow):         MainWindow.setObjectName("MainWindow")         MainWindow.resize(540, 630) #设置显示大小         sizePolicy = QSizePolicy(QSizePolicy.Expanding, QSizePolicy.Preferred)         sizePolicy.setHorizontalStretch(0)         sizePolicy.setVerticalStretch(0)         sizePolicy.setHeightForWidth(MainWindow.sizePolicy().hasHeightForWidth())         MainWindow.setSizePolicy(sizePolicy)         # MainWindow.setMinimumSize(QSize(540, 630)) #x*y最小规则         icon = QIcon() #加载ico图标         icon.addFile(u"Trans.ico", QSize(), QIcon.Normal, QIcon.Off)         MainWindow.setWindowIcon(icon)           self.centralwidget = QWidget(MainWindow)         self.centralwidget.setObjectName(u"centralwidget")           # 说明         self.pushButton = QPushButton(self.centralwidget)         self.pushButton.setObjectName(u"pushButton")         self.pushButton.setGeometry(QRect(400, 320, 100, 40))         # 开始         self.pushButton_2 = QPushButton(self.centralwidget)         self.pushButton_2.setObjectName(u"pushButton_2")         self.pushButton_2.setGeometry(QRect(400, 370, 100, 40)) #xy位置-大小         self.pushButton_2.setEnabled(False)  #初始化不可点击         # 打开结果         self.pushButton_5 = QPushButton(self.centralwidget)         self.pushButton_5.setObjectName(u"pushButton_2")         self.pushButton_5.setGeometry(QRect(400, 420, 100, 40)) #xy位置-大小         self.pushButton_5.setEnabled(False)  #初始化不可点击         # 出错列表         self.pushButton_6 = QPushButton(self.centralwidget)         self.pushButton_6.setObjectName(u"pushButton_2")         self.pushButton_6.setGeometry(QRect(400, 470, 100, 40)) #xy位置-大小         self.pushButton_6.setEnabled(False)  #初始化不可点击         # 进度条         self.progressBar = QProgressBar(self.centralwidget)         self.progressBar.setObjectName(u"progressBar")         self.progressBar.setGeometry(QRect(40, 550, 470, 30))         self.progressBar.setStyleSheet(u"\n"                         "QProgressBar::chunk\n"                         "{\n"                         "border-radius:15px;\n"                         "background:qlineargradient(spread:pad,x1:0,y1:0,x2:1,y2:0,stop:0 #70649A,stop:1  #26B4FF);\n"                         "border-radius:4px;/*\u6591\u9a6c\u7ebf \u5706\u89d2*/\n"                         "border:1px solid black;\n"                         "background-color:skyblue;\n"                         "width:6px;margin:0.5px;/*\u5bbd\u5ea6\u548c\u5916\u8fb9\u8ddd*/\n"                         " \n"                         "}\n"                         "QProgressBar#progressBar\n"                         "{\n"                         "height:20px;\n"                         "text-align:center;\n"                         "font-size:14px;\n"                         "color:white;\n"                         "border-radius:11px;\n"                         "background: #0f1423 ;\n"                         "}")         self.progressBar.setValue(0)         # 操作栏==============================================         self.groupBox = QGroupBox(self.centralwidget)         self.groupBox.setObjectName(u"groupBox")         self.groupBox.setGeometry(QRect(40, 20, 470, 100))         # 新建按钮组         self.buttonGroup = QButtonGroup(MainWindow)         self.buttonGroup.setObjectName(u"buttonGroup")         # 汇总         self.radioButton = QRadioButton(self.groupBox)         self.radioButton.setChecked(True)  #默认选中         self.radioButton.setObjectName(u"radioButton")         self.radioButton.setGeometry(QRect(50, 30, 185, 50))         # 分类         self.radioButton_2 = QRadioButton(self.groupBox)         self.radioButton_2.setObjectName(u"radioButton_2")         self.radioButton_2.setGeometry(QRect(235, 30, 185, 50))         # 添加按钮到按钮组         self.buttonGroup.addButton(self.radioButton)         self.buttonGroup.addButton(self.radioButton_2)         # 操作栏==============================================           # 路径栏==============================================         self.groupBox_2 = QGroupBox(self.centralwidget)         self.groupBox_2.setObjectName(u"groupBox_2")         self.groupBox_2.setGeometry(QRect(40, 140, 470, 150))         # 文件、模板         self.comboBox = QComboBox(self.groupBox_2)         self.comboBox.setObjectName(u"comboBox")         self.comboBox.setGeometry(QRect(20, 30, 350, 40))         self.comboBox_2 = QComboBox(self.groupBox_2)         self.comboBox_2.setObjectName(u"comboBox_2")         self.comboBox_2.setGeometry(QRect(20, 90, 350, 40))         self.pushButton_3 = QPushButton(self.groupBox_2)         self.pushButton_3.setObjectName(u"pushButton")         self.pushButton_3.setGeometry(QRect(380, 30, 70, 40))         self.pushButton_4 = QPushButton(self.groupBox_2)         self.pushButton_4.setObjectName(u"pushButton")         self.pushButton_4.setGeometry(QRect(380, 90, 70, 40))         # 路径栏==============================================           # 其他栏==============================================         self.groupBox_3 = QGroupBox(self.centralwidget)         self.groupBox_3.setObjectName(u"groupBox_3")         self.groupBox_3.setGeometry(QRect(40, 310, 300, 210))         # 首列添加序号、数字前加单引号、保存结果等         self.checkBox = QCheckBox(self.groupBox_3)         self.checkBox.setObjectName(u"checkBox")         self.checkBox.setGeometry(QRect(30, 30, 250, 40))         self.checkBox_2 = QCheckBox(self.groupBox_3)         self.checkBox_2.setObjectName(u"checkBox_2")         self.checkBox_2.setGeometry(QRect(30, 60, 250, 40))         self.checkBox_3 = QCheckBox(self.groupBox_3)         self.checkBox_3.setChecked(True)  #默认选中         self.checkBox_3.setObjectName(u"checkBox_3")         self.checkBox_3.setGeometry(QRect(30, 90, 250, 40))         self.checkBox_4 = QCheckBox(self.groupBox_3)         self.checkBox_4.setChecked(True)  #默认选中         self.checkBox_4.setObjectName(u"checkBox_3")         self.checkBox_4.setGeometry(QRect(30, 120, 250, 40))         self.checkBox_5 = QCheckBox(self.groupBox_3)         self.checkBox_5.setObjectName(u"checkBox_3")         self.checkBox_5.setGeometry(QRect(30, 150, 250, 40))           # 其他栏==============================================           MainWindow.setCentralWidget(self.centralwidget)         self.statusbar = QStatusBar(MainWindow)         self.statusbar.setObjectName(u"statusbar")         MainWindow.setStatusBar(self.statusbar)         self.retranslateUi(MainWindow)         QMetaObject.connectSlotsByName(MainWindow)         def retranslateUi(self, MainWindow):         MainWindow.setWindowTitle(QCoreApplication.translate("MainWindow", u"Doc2Xlsx", None))         self.pushButton_2.setText(QCoreApplication.translate("MainWindow", u"开始运行", None))         self.groupBox.setTitle(QCoreApplication.translate("MainWindow", u"操作", None))         self.radioButton.setText(QCoreApplication.translate("MainWindow", u"汇总", None))         self.radioButton_2.setText(QCoreApplication.translate("MainWindow", u"分类", None))         self.pushButton.setText(QCoreApplication.translate("MainWindow", u"相关说明", None))         self.groupBox_3.setTitle(QCoreApplication.translate("MainWindow", u"其他", None))         self.checkBox.setText(QCoreApplication.translate("MainWindow", u"首列添加序号", None))         self.checkBox_2.setText(QCoreApplication.translate("MainWindow", u"数字前添加单引号", None))         self.checkBox_3.setText(QCoreApplication.translate("MainWindow", u"保存结果文件", None))         self.checkBox_4.setText(QCoreApplication.translate("MainWindow", u"Excel后台运行", None))         self.checkBox_5.setText(QCoreApplication.translate("MainWindow", u"对日期格式化", None))         self.groupBox_2.setTitle(QCoreApplication.translate("MainWindow", u"路径", None))         self.pushButton_3.setText(QCoreApplication.translate("MainWindow", u"文件", None))         self.pushButton_4.setText(QCoreApplication.translate("MainWindow", u"模板", None))         self.pushButton_5.setText(QCoreApplication.translate("MainWindow", u"打开结果", None))         self.pushButton_6.setText(QCoreApplication.translate("MainWindow", u"出错列表", None))

然后是引用它的主功能文件，如果不用图形界面只要修改功能的话，只需要看下面的就好了：

[Python] 纯文本查看 复制代码

?

001 002 003 004 005 006 007 008 009 010 011 012 013 014 015 016 017 018 019 020 021 022 023 024 025 026 027 028 029 030 031 032 033 034 035 036 037 038 039 040 041 042 043 044 045 046 047 048 049 050 051 052 053 054 055 056 057 058 059 060 061 062 063 064 065 066 067 068 069 070 071 072 073 074 075 076 077 078 079 080 081 082 083 084 085 086 087 088 089 090 091 092 093 094 095 096 097 098 099 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342

# -*- coding: UTF-8 -*- import win32com.client as wc import os, re, glob, time, shutil from bs4 import BeautifulSoup from pydocx import PyDocX from PySide2.QtWidgets import * from PySide2.QtCore import Slot import D2XUI   class Doc2Xlsx(QMainWindow, D2XUI.d2xUI):       def __init__(self):         super(Doc2Xlsx, self).__init__()         self.addSerial = False         self.addQuotaMark = False         self.saveResult = True         self.excelHidden = True         self.regularDate = False         self.dateLine = []         self.pattern = re.compile(r"(\d{2,4})\D*(\d{0,2})\D*(\d{0,2})")         self.operate = "summary"         self.wordPath = ""         self.modelPath = ""         self.openResult = ""           # 引用D2XUI函数         self.setupUi(self)         # 操作选项         self.buttonGroup.buttonClicked.connect(self.Operate)         # 路径选项         self.pushButton_3.clicked.connect(self.FilePath)         self.comboBox.currentIndexChanged.connect(self.FilePathRe)         self.pushButton_4.clicked.connect(self.ModelPath)         self.comboBox_2.currentIndexChanged.connect(self.ModelPathRe)         # 其他选项         self.checkBox.clicked.connect(self.Others)         self.checkBox_2.clicked.connect(self.Others)         self.checkBox_3.clicked.connect(self.ResHidden)         self.checkBox_4.clicked.connect(self.ResHidden)         self.checkBox_5.clicked.connect(self.RegularDate)         # 说明         self.pushButton.clicked.connect(self.Explain)         # 开始         self.pushButton_2.clicked.connect(self.Start)         # 结果文件         self.pushButton_5.clicked.connect(self.Result)         # 出错列表         self.pushButton_6.clicked.connect(self.Errorlist)         @Slot() #开始运行     def Start(self):         self.progressBar.reset()         # doc转docx并删除doc         docs = [i for i in glob.glob(self.wordPath + "/*.doc")]         for doc in docs:             self.doc2docx(doc)             self.delDoc(doc)           if self.operate == "summary":             self.docx2xlsx(self.wordPath, self.modelPath)         elif self.operate == "classify":             self.docxCategory(self.wordPath)       @Slot() #选择汇总或分类     def Operate(self):         if self.buttonGroup.checkedButton().text() == "汇总":             self.operate = "summary"             self.pushButton_4.setEnabled(True) #设置模板可选中状态         else:             self.operate = "classify"             self.pushButton_4.setEnabled(False)         self.changeStart()         self.progressBar.setValue(0)       @Slot() #选择路径     def FilePath(self):         self.wordPath = QFileDialog.getExistingDirectory(self, "选择Word文件所在文件夹")         self.comboBox.addItem(self.wordPath)         self.comboBox.setCurrentText(self.wordPath)         self.changeStart()         self.progressBar.setValue(0)       @Slot() #通过combobox选择历史路径     def FilePathRe(self):         self.wordPath = self.comboBox.currentText()         self.progressBar.setValue(0)       @Slot()     def ModelPath(self): #选择模板         self.modelPath = QFileDialog.getOpenFileName(self, "选择模板文件","","Docx文件(*.docx)")[0]         self.comboBox_2.addItem(self.modelPath)         self.comboBox_2.setCurrentText(self.modelPath)         self.changeStart()         self.progressBar.setValue(0)       @Slot() #通过combobox选择历史模板     def ModelPathRe(self):         self.modelPath = self.comboBox_2.currentText()         self.progressBar.setValue(0)          @Slot() #其他：首列加序号、数字加单引     def Others(self):         self.addSerial = True if self.checkBox.isChecked() else False         self.addQuotaMark = True if self.checkBox_2.isChecked() else False         self.progressBar.setValue(0)       @Slot() #其他：日期格式化     def RegularDate(self):         if self.checkBox_5.isChecked():             self.regularDate = True             dline, okPressed = QInputDialog.getMultiLineText(self, "日期格式化","日期所在Excel表列数(多列换行表示):")             if okPressed:                 if not dline:                     self.checkBox_5.setChecked(False)                     self.regularDate = False                 else:                     self.dateLine = [int(i.strip()) for i in dline.split("\n")]             else:                 self.regularDate = False                 self.checkBox_5.setChecked(False)         else:             self.regularDate = False         self.progressBar.setValue(0)       @Slot() #其他：保存结果、Excel隐藏     def ResHidden(self):         if self.checkBox_3.isChecked():             self.saveResult = True             self.checkBox_4.setEnabled(True)             self.excelHidden = True if self.checkBox_4.isChecked() else False         else:             self.saveResult = False             self.checkBox_4.setChecked(False)             self.checkBox_4.setEnabled(False)             self.excelHidden = False         self.progressBar.setValue(0)       @Slot() #打开结果文件     def Result(self):         if self.operate == "classify":             try:                 os.startfile(self.wordPath + "/_cate/")             except:                 pass         elif self.operate == "summary" and self.saveResult:             try:                 os.startfile(self.wordPath + "/result.xlsx")             except:                 pass         self.progressBar.setValue(0)       @Slot() #相关说明     def Explain(self):         msgBox = QMessageBox()         # msgBox.setIcon(QMessageBox.Warning)         msgBox.setWindowTitle ("说明")         msgBox.addButton ("我已知晓=。=", QMessageBox.AcceptRole)         msgBox.setStyleSheet('''             QMessageBox { width:140px;height:20px;text-align:center;font-size:20px; };             QPushButton { background-color: rgb(255, 85, 0); color: rgb(85, 255, 0); }''')         message = '''         使用前请仔细阅读以下说明:         0. 注意数据备份;         1. 表格无变化或提取内容位于变化部分前则无需分类;         2. 分类后每个类别需分别制作模板文件进行汇总;         3. [模板]:在相应单元格填Excel列数并删除所有内容;         4. [数字前加单引号]:可避免手机号等科学记数法;         5. [首列添加序号]:首列为序号而非数据可勾选;         6. [保存结果文件]:后台运行保存否则前台显示结果;         7. [对日期格式化]:统一为2021.01.01格式;         8. [Excel后台运行]:不显示Excel操作窗口;         9. doc会转换为docx并删除原doc;         10. 数据出错与否最好都手动核查.'''         msgBox.setText(message)         msgBox.exec_()       @Slot() #打开出错列表     def Errorlist(self):         try:             os.startfile(self.wordPath + "/error.txt")         except:             self.pushButton_6.setEnabled(False)         self.progressBar.setValue(0)               @Slot() #切换操作时按钮变化     def changeStart(self):         if self.operate == "classify" and self.wordPath:             self.pushButton_2.setEnabled(True)         elif self.operate == "summary" and self.wordPath and self.modelPath:             self.pushButton_2.setEnabled(True)         elif self.operate == "summary" and self.wordPath and not self.modelPath:             self.pushButton_2.setEnabled(False)         self.progressBar.setValue(0)         # 以下是功能函数=======================================================     # doc2docx     def doc2docx(self, file):         word = wc.Dispatch("Word.Application")         doc = word.Documents.Open(file)         doc.SaveAs(file + "x", 12)  #12代表转换后为docx文件         doc.Close         word.Quit()   #结束后台进程         # Del doc     def delDoc(self, file):         time_start=time.time()         while True:             time_end=time.time()             wmi = wc.GetObject("winmgmts:")             proCodeCov = wmi.ExecQuery('select * from Win32_Process where Name="WINWORD.EXE"')             if len(proCodeCov) <= 0:                 os.remove(file)                 return             time.sleep(3)             if time_end - time_start > 20:                 return         # 是否为数字,身份证号及手机号等     def isNumber(self, s):         try:             # float(s)             int(s)             return True         except ValueError:             return False         # 根据html tr&td 标签分类     def docxCategory(self, path):         docxs = [i for i in glob.glob(path + "/*.docx")]         self.progressBar.setRange(0, len(docxs) + 1)         self.progressBar.setValue(0)         docx_cate = {}           for c, docx in enumerate(docxs):             self.progressBar.setValue(c + 2)             html = PyDocX.to_html(docx)             soup = BeautifulSoup(html, 'html.parser')             elem = soup.find_all("tr")                              #每个tr标签对应表格的一行                                  elem1 = re.sub(r"(?<=<td).+?(?=</td>)", "", str(elem))  #去除单元格标签td之间的内容               if elem1 in docx_cate.keys():                 e = docx_cate[elem1]                 e.append(docx)                 docx_cate[elem1] = e             else:                 docx_cate[elem1] = [docx]         self.progressBar.setValue(0)           flag = 1         for c, docx in enumerate(docx_cate.values()):             path_cate = path + "/_cate/类别%d/"%(c+1)             if not os.path.exists(path_cate):                 os.makedirs(path_cate)             for i in docx:                 shutil.copyfile(i, path_cate + os.path.split(i)[1]) #复制,移动则shutil.move                 self.progressBar.setValue(flag + 1)                 flag += 1         self.pushButton_5.setEnabled(True) #打开结果按钮                 # docx2xlsx     def docx2xlsx(self, path, model):         if os.path.exists(path + "/error.txt"): #删除上一次的出错列表             os.remove(path + "/error.txt")         docxs = [i.replace("\\", "/") for i in glob.glob(path + "/*.docx")]         if model in docxs:             docxs.remove(model)           self.progressBar.setRange(0, len(docxs) + 1)         self.progressBar.setValue(0)                   # 获取模板文件所需数据的位置         html = PyDocX.to_html(model)         soup = BeautifulSoup(html, 'html.parser')         elem = soup.find_all("tr")         data = {}         for row, tr in enumerate(elem):             if tr.text:                           #该行有需要的数据                 val = {}                 units = tr.find_all("td")         #获取需要的单元格序号                 for unit, td in enumerate(units):                     if td.text:                         val[unit] = td.text                 data[row] = val                   # 创建xlsx         xl = wc.Dispatch('Excel.Application')         if self.saveResult: #是否可见             xl.Visible = False if self.excelHidden else True         else:             xl.Visible = True         xl.Workbooks.Add()         xlBook = xl.Workbooks(1)         xlSheet = xl.Sheets(1)                   # 数据汇总         for c, docx in enumerate(docxs):             html = PyDocX.to_html(docx)             soup = BeautifulSoup(html, 'html.parser')             elem = soup.find_all("tr")             if self.addSerial:                 xlSheet.Cells(c+1, 1).Value = c + 1             try:                 for r in data.keys():                     for u in data[r].keys():                         text = elem[r].find_all("td")[int(u)].text.strip()                         if self.regularDate and int(data[r][u]) in self.dateLine: #格式化日期                             try:                                 date = self.pattern.findall(text)[0]                                 text = "{}.{:0>2s}.{:0>2s}".format(date[0], date[1], date[2]) if date[2] else "{}.{:0>2s}".format(date[0], date[1])                             except:                                 pass                         if self.addQuotaMark and self.isNumber(text): #数字前加单引号                             text = "'" + text                         xlSheet.Cells(c+1, int(data[r][u])).Value = text             except Exception as e:                 self.pushButton_6.setEnabled(True) #设置按钮为可点击状态                 with open(path + "/error.txt", "a", encoding="utf-8") as g:                     g.write(str(os.path.split(docx)[1]) + "\n")             self.progressBar.setValue(c + 2)                       if self.saveResult:             xlBook.SaveAs(path.replace(r"/", "\\") + "\\result.xlsx")  #实际单\才能正常保存             self.pushButton_5.setEnabled(True) #打开结果按钮             if self.excelHidden:                 xl.Quit() #关闭ark     def main():     app = QApplication([])     d2x = Doc2Xlsx()     d2x.show()     app.exec_()     if __name__ == '__main__':     main()

有评论说Excel催化剂能做到，我去下载了不知道怎么用功能是在太多了，有需要的朋友可以下载。
还有评论要的VBA版我也传上来了，我记得有官网的，程序里面应该就有，出错可以去看一下。

【链接】
https://pan.lanzouj.com/b00u7yupc
密码：9idc


https://wwa.lanzouj.com/b00u8190j
密码:bgm9




连接打不开的，蓝奏问题，把lanzous改为lanzoui。

hxw139

lijt16 发表于 2021-2-2 19:35
word表格汇总到excel，很难理解？

额   懂起了  不好意思  平时都是excel表格到word  这个还没操作过  感谢分享

lijt16

eersoft 发表于 2021-12-2 21:45
楼主分享的vba版是我的，楼主不错原样分享了。前段时间看到某知名大V 王D田，在视频中将我的工具 ...

牛，我是在做的过程中不经意发现你的工具的，真的厉害

blueczw

不明白这是嘛意思

lijt16

blueczw 发表于 2021-2-2 19:01
不明白这是嘛意思

emmmm你这。。

hxw139

干嘛用的啊？不明所以

lijt16

hxw139 发表于 2021-2-2 19:33
干嘛用的啊？不明所以

word表格汇总到excel，很难理解？

ciker_li

感谢楼主分享源码

bsjasd

多谢楼主分享

hebeijianke

感谢楼主分享，向楼主学习

Wapj_Wolf

近700行的代码看着有些眼晕，够我消化一阵子的了，谢谢大佬无私分享……